T. H. Vanderspurt, S. C. Emerson, T. Davis, J. MacLeod, G. Marigliani, A. Peles, S. Seiser, Y. She, & R. Willigan

United Technologies Research Center, East Hartford, CT

20 May 2009

A Novel Slurry Based Biomass Reforming Process

(DE-FG36-05GO15042)

This document does not contain any proprietary, confidential, or otherwise restricted information.

Project ID #PD_28_Vanderspurt

2

Overview

TimelineStart: May 2005End: August 201060% Complete

BudgetTotal Project Funding

DOE share: $3MContractor share: $750k

Funding Received in FY08$800k

Funding for FY09$0k

BarriersT. Capital Costs and Efficiency of Biomass Gasification/Pyrolysis Technology

PartnersUniversity of North Dakota Environment Energy Research Center (UND-EERC)

Hydrolysis experimental studiesSlurry characterizationCost analysisProof-of-concept demonstration

3

Target Current Status 2008 Status

Hydrogen Cost (Plant Gate)a 1.60 $/gge $1.27/kg H2 ($0.95/kg – $1.85/kg) $1.58/kg H2 – $2.16/kg H2

Total Capital Investment b $150M $177M ($71M – $365M) $203M – $291M

Energy Efficiency c 43% 53.9% (53.9% – 58%) 46.6%

a. Gallon of gasoline equivalent (gge) ≈ kg H2b. Assuming 300 psig delivery pressurec. Plant H2 Efficiency

Development of an initial reactor and system design, with cost projections, for a biomass slurry hydrolysis and reforming process for H2 production

Efficiency & H2 cost exceed gasification targetsCapital costs dependent on H2 delivery pressure & H2 separation membrane

Development of cost effective catalysts for liquid phase reforming of biomass hydrolysis-derived oxygenates

Catalyst based on commercially scaled-up materialProof-of-concept demonstration of a micro-scale pilot system based on liquid phase reforming of biomass hydrolysis-derived oxygenates

Phase II work not yet started

Project Objectives & DOE Target Status

4

H2 Compressor Is Best Choice for Delivery of H2 at 300 psia atmH2 flux, SCFH/ft2

Membrane cost, $/ft2

Capital cost,$ million

Membrane cost in capital, %

Efficiency, %

H2 selling price,$/kg

H2production,kg/day

Compressor

60 1,500 365 73.8 53.9 1.85 2.65x105

200 1,500 177 45.7 53.9 1.27 2.65x105

200 100 101 5.3 53.9 1.04 2.65x105

H2 is delivered at 300 psia pressure

Technical Accomplishments: Economic Analysis

5

Approach: Biomass Slurry to Hydrogen Concept

Fuel flexible, using raw, ground biomass such as wood or switch grass Carbon neutral means to produce HydrogenH2 separation: Leverage experience with Advanced Pd membranes

Alternative: Biomass Gasification to Hydrogen Concept

Fuel flexible, using raw, ground biomass such as wood or switch grass Gasification & PSA: Leverage existing technologiesProcess complexity

Gasification Reforming Shift PSA

H2

Char combustor &cyclones

Sand Char & sand

Steam

Air

Slurry of ≈10 %Ground Biomass(Wood)

0.09-0.13 kg H2/kg Biomass (dry)

6

Technical Accomplishments: 2008 SummaryExplored glycerol liquid phase reforming with atomistic modelingModified experimental setup to ensure measurement of true hydrogen production rates

Semi-batch reactors used for screeningScreened several catalyst candidates, including a commercially available Pt/Al2O3 as a baseline comparison

Downselected a ceria-zirconia based catalystExamined reforming in the presence of acid (sulfur) and base

Performed material dissolution and char formation studies Conducted biomass hydrolysis studiesChar formation prevented adequate catalyst evaluationConcluded that alkali reforming avoided catalyst poisoning & side reactions that form char

Proof-of-concept demonstrations of complete reforming of wood and wood components in alkali system

7

Simplified Diagram for Alkaline Reforming Plant

Ref

orm

er

Pd

mem

bran

e

8

Technical Accomplishments: System Modeling

Adjusted feedstock assumption from previous modelNow 72.6% cellulose; 27% lignin surrogate (C8H8O3); 0.4% ash

Reduced hydrolysis to one-step alkali processAlkali is derived from potassium in woodComplete hydrolysis of lignin

Eliminated sulfur recoveryCombust process gases instead of unconverted ligninReduced reformer size based on ethanol kinetics

PretreatT110

Alkaline

130Hyd201

Biom110

P210

Refo rm

300

H2O120

Exp411

HEX411

Burner500

Air140

H2Storage

B140

P120

P130

TankT111

P110

Pdme

mb40

0

Exp410

HEX140

HEX110

Sep310

Burner500

Decanter

H2O150 P150

HEX410Hot water

HEX111Exp310

Hyd202

S410

P410

Fil310

P310

Fil510

SO2 Oxid

B510

SO3 Scruber

Exhaust

AshGas

Discharge

HEX510Steam&Water

Hot water

TEE310

Wastewater treatment plant

gone

gone

HYSYS Plant Design for Alkali Hydrolysis: Wood H2, CH4, C2H6 + C3H8

9

Technical Accomplishments: Economic Analysis

UTRC task of delivering H2 at 300 psia requires either back pressure on membrane or compressor system, but compressor is more cost effectiveAt DOE H2 membrane targets, H2 price is ($1.18–$1.27)/kg H2Reduced flux membranes yield a H2 price of ($1.76 –$1.85)/kg H2UTRC advanced concept membrane gives a H2 price of ($0.95 –$1.04)/kg H2

Revised Economic Projections for Alkali HydrolysisH2 delivery options

Membrane options Membrane Flux /

[SCFH/ft2]

Membrane Cost / [$/ft2]

Membrane % of

capital cost

Efficiency /%

H2 selling price /

[$/kg H2]

1 atm DOE targets 200 1500 54.6 58.1 1.18

DOE cost, reduced flux 60 1500 80.0 58.1 1.76

UTRC advanced concept 200 100 7.4 58.1 0.95

150 psia (compressor)

DOE targets 200 1500 45.7 54.9 1.27

DOE cost, reduced flux 60 1500 73.8 54.9 1.85

UTRC advanced concept 200 100 5.3 54.9 1.04

300 psia (compressor)

DOE targets 200 1500 45.7 53.9 1.27

DOE cost, reduced flux 60 1500 73.8 53.9 1.85

UTRC advanced concept 200 100 5.3 53.9 1.04

10

Semi-Batch Reactor

Titanium Reactor

Maximum Operating Conditions:Pressure (2000 psig)

Temperature (310 °C)

NeodymiumStir Plate

Cooling Water In

Cooling Water Out

SmCo Ti Sheathed Stir Bar

Catalyst Suspended in Ti Basket

+ -

+ -

N2

IR Furnace

N2 and Product Gas

Tc

Biomass

Aqueous Solution

(Not to Scale)

11

Technical Accomplishments: Total Reforming of Yellow Poplar

Semi-batch conversion of yellow poplar to H2, CH4, C2H6, & C3H8

For LHVwood = 18 kJ/g; If all H2 is from reforming, H2 LHV efficiency would be 75% Complete conversion of wood, including lignin: only trace organic carbon in liquidAlkane byproducts could provide enough energy to run endothermic reformer

≈100% conversion of yellow poplar to gas with continuous H2 removal

TitaniumSemi-BatchReactor

Effluent product profiles from the hydrolysis and liquid phase reforming of 1 wt%yellow poplar at 310 °C in 0.1M K2CO3 with a 0.5 L/min N2 sweep gas.

Reaction product liquor. Titanium corrosion resulted in potassium titanate formation as well as some H2 production.

12

Nano-engineered Catalyst

HemicelluloseCellulose

LigninOthers

OH- + nH2O

Lignin + Fragments

Reacted Glucose Reacted C5 sugarsUronic Acids

Tar and Solids H2 + CO2 +CH4 + C2H6 + C3H8

+OH- + nH2O

Alkyl & methoxy phenols etc.

Heat

Technical Accomplishments: Proposed Wood Reforming Pathway

Unreformed hydrolysis products form intractable tars over time.

PeelingReaction

13

Technical Accomplishments: Ethanol Reforming

Develop mathematical model based on simple feedsExtend model to more complicated feeds like wood via a surrogate compound based on compositional analysis

For example, yellow poplar might be represented by C6H8O4Use models to understand H2 selectivity, reactor sizing, and performance

Reforming Reaction Network Determined from Ethanol ExperimentsGeneric Reaction Network

Ethanol Reaction Network

14

Technical Accomplishments: Ethanol Reforming

Preliminary flow reactor tests confirmed model for ethanol systemEthanol system performance understoodReforming to H2 LHV efficiency ≈52.7% for a 1.7 wt% ethanol feedGoal is to extend the model to wood feedstocks

Developed Kinetic Model for Reactor Sizing and Economics

Ethanol reforming at 310 °C compared to model

15

Technical Accomplishments: Yellow Poplar Flow Reactor

First flow reactor tests with hydrolyzed woodMuch higher H2/CH4 yields than with ethanolReactor packing material deteriorated, causing apparent activity loss:New Packing IdentifiedHydrolyzed wood liquor is unstable, darkens rapidly and develops solids, may require process design change.

Beginning to Understand the Flow Operation & Kinetics of Wood Reforming

16

Collaborators

University of North Dakota Energy and Environmental Research Center

Subcontractor for hydrolysis studies & demonstrationContract signed in December 2008Secured of hybrid poplar for UTRC & UND studiesEffort currently focused on characterizing components of alkali hydrolyzed biomass2009 objective to examine the effect of several variables (e.g., base concentration, temperature, and pressure) on hydrolysis products and conversion

Catalyst oxide support technology was transferred to a vendor

Membrane Technology Development on DE-FC26-07NT43055Power+Energy (Industry)

Manufacture of hydrogen separatorsUTRC alloy fabrication

Metal Hydride Technologies (Ted Flanagan from Univ. of Vermont)Fundamental experiments on hydrogen solubilityExperimental measurements of alloy systems for thermodynamic phase modeling

17

Proposed Future Work

2009Evaluate kinetic model on wood reformingComplete techno-economic analysisDemonstrate reformer operation with a H2 separatorComplete UND hydrolysis study

Future: Phase 2Develop 1-kWe demonstration system

18

Summary

Provide overview of project progressDemonstrated >97% conversion of wood to hydrogen with high efficiencyDeveloping kinetic understanding through model compounds (e.g., ethanol)Commenced Flow reactor experiments to extend kinetic model to woodSystem efficiency and economics favorable compared to gasification

GO/NO GO decision: Received a GO decision in March 2009

Future work: Reactor Design, Base Treatment Protocol, catalyst morphology and catalyst atomic/electronic structure, reformer+membrane demonstration

Work performed under the DOE Grant DE-FG36-05GO15042 was authorized in part under a research license for the Aqueous Phase Reforming Process (Patent 6,699,457; 6,694,757; 6,694,758 [and all other licensed issued patents]) from Virent Energy Systems in Madison, Wisconsin.

19

Supplemental Slides

Supplemental slides follow

20

Modeling and Simulation Assumptions

Heat losses: 4–5% of the heat load on heat exchangers.

Energy consumption for biomass pretreatment: 1.8% of the biomass feed LHV.

Pd alloy membrane selectivity of H2: 100%

The biomass feedstock: 72.6% cellulose, 27% lignin surrogate (C8H8O3) and 0.4% ash.

Biomass feedstock price is: $0.05/kg (dry basis; twice the value in the H2A tool)

One-stage hydrolysis processes operated at temperature of 310 °C.

95% conversion of cellulose was assumed based on our preliminary experimental results.

In the baseline design, the yield of H2 and CO2 from glucose: 90% from sugar hydrolyzed from cellulose and 95% from lignin surrogate over the advanced high activity, high selectivity catalyst system.

Ultra-pure H2 from membrane compressed to 300 psia for pipeline

24

Act

ivity

Approach: UTRC Catalyst Discovery ParadigmSuccessfully Employed to Develop High Activity, Long Lived (>5X) Catalysts

Characterization

High active surface area Large pore Nanocrystalline structure

~100% NM dispersion

Conceptual Catalyst Design Quantum Mechanical Atomistic Modeling for advanced catalysts

design

Catalyst Synthesis

⇔ ⇔ ⇔

1000 2000 30004000Wavenumbers (cm-1)

Catalyst A

Catalyst B =Kinetic Expressions Derived

From Reaction Data

Superior Performance

⇔ Act

ivity

Unfavorable

• Literature Search:Catalyst Chemistry

• Nano-Engineering Specifications: - Pore Sizes, - Porous Particle Sizes

25

Targeted 1-kWe Demonstration System

Combine technology advances from Biomass to H2 program with advanced high pressure H2 membrane concept

Operation at T ≈ 310 °C, P > 1417 psig1 kWe H2 requires 12.1 L/min H2 (2 kW H2)

5.4 L/h (90 mL/min) of biomass slurry (≈10 wt% biomass)Reactor volume <0.5 L

Energy required only for liquid pumping and heating of membrane reactor

Combine liquid phase reforming with advanced H2 separation membrane

26

History- Overall Scheme

Initial plan was aqueous phase reforming of biomass in acidUse acid to hydrolyze biomassDevelop novel sulfur tolerant catalyst to reform hydrolyzed species

Numerous technical hurdles, including:Sulfur poisoning, catalyst dissolution, support dissolution, hydrolysis optimization.

Attempts to create sulfur tolerant catalyst unsuccessfulActivities in catalyst development further deterred by discovery of charring issues in acid hydrolysis of real biomassGlycerol testing indicated acid catalyzed polymerization

Course changed to use of base:Similar technical challenges, catalyst competition, corrosive environment, little literatureHowever, there exists a path for hydrolysis

Have proven feasibility of system Complete conversion of raw biomass to gas at 310°C with >70% selectivity to H2.

27

Potassium Species

Questions regarding which base to useKOH pH 14K2CO3 pH 10.5KHCO3 pH 8.5

K2CO3 chosen for studyBehaves like KOH at high temperaturesRecyclable

Issues closing carbon balanceCO2 Scrubbing

28

Reforming with Wood Components

Cellulose, Hemicellulose, Lignin comprise ~43.9,15.7, 26.0 wt% wood1% Feed in 0.2M KOH at 310°C, 110atm, 2% Pt-3% Re/Ce0.55Zr0.45O2

•Complete LPR of Cellulose and Xylan in 0.2M KOH•H2 Selectivity >80%•Increased temperature led to increased conversion

29

LPR with Wood Components- Catalyst Stability

1% Cellulose in 0.2M KOH at 310°C, 110atm, 2% Pt-3% Re/Ce0.55Zr0.45O2

Repeated 3X, no apparent loss in activity

Time (Minutes)

0 200 400 600 800 1000

Hyd

roge

n G

ener

atio

n, m

ol/m

in

0.00000

0.00005

0.00010

0.00015

0.00020

0.00025

Fresh Catalyst2X Used Catalyst

30

Pre-Hydrolysis vs. In-Situ Hydrolysis

1% Yellow Poplar at 310°CMatched conversionMatched selectivity

Prehydrolysis In-Situ